Uses Oriented Bounding Boxes (OBB) to handle the high aspect ratios of vessels.
Future work involves improving detection in adverse weather conditions. ⚓ HRSC2016 is a gold standard for ship detection research.
Deep models effectively distinguish ships from dock structures but struggle with very small vessels in low-contrast conditions. 5. Conclusion Summary of model performance on HRSC2016. sc24016-VTv08731.rar
Includes ships in open sea, docked in ports, and near coastal clutter. 3. Methodology
Image resizing and normalization to match model requirements. Model Selection: YOLOv5/v8: For real-time detection performance. Faster R-CNN: For high-precision localization. Rotated Det: Specifically designed for OBB detection tasks. Uses Oriented Bounding Boxes (OBB) to handle the
Evaluated using Mean Average Precision (mAP) and F1-score.
Benchmark deep learning architectures on the HRSC2016 dataset . 2. Dataset Overview Includes ships in open sea, docked in ports,
Oriented Bounding Boxes (OBB) are due to the long, narrow shape of ships.